|
Plorkyeran posted:you can't meaningfully implement a b+ tree in a language that doesn't give you any control over memory layout. you can write all of the algorithms and such, but if the end result doesn't have your leaves in contiguous blocks of memory it's all sort of pointless i'm sure some fucker can and has done something like this with typed arrays
|
# ? Nov 15, 2016 02:03 |
|
|
# ? May 9, 2024 21:45 |
|
eschaton posted:you totally have control over memory layout it's disgusting but it works well enough to run dosbox and mame in the browser (!!!)
|
# ? Nov 15, 2016 02:04 |
|
typescript is meant to be a superset of JavaScript to enable gradual migration and conversion of existing projects. it's a good language that has a lot of use in the Real World
|
# ? Nov 15, 2016 02:41 |
|
hackbunny posted:it's disgusting but it works well enough to run dosbox and mame in the browser (!!!) ive written a compiler that compiles a subset of c to a mips stack machine, so it can't be any more perilous than that. i guess that's another project to go on my pile of poo poo to do when im not completely miserable.
|
# ? Nov 15, 2016 02:45 |
|
eschaton posted:you totally have control over memory layout just target asm.js, op
|
# ? Nov 15, 2016 03:41 |
|
every time i read "asm.js" i think "rear end to mouth dot js" to myself for some reason
|
# ? Nov 15, 2016 03:59 |
|
lol, now ive been thinking about this, and i realize that i think the various books and guides on mumps out there are wrong-ish. everything describes the structure of globals and arrays as a b+ tree, but i think it's important to call out that it's really a b+ tree where any leaf can be a pointer to another b+ tree. afterall, a b+ tree is only an efficient means to represent a data structure like: code:
code:
code:
code:
i suppose the other way it could be implemented underneath would be to treat the entire comma-delimited series of keys as a single composite key (and this might explain cache's limitation of 511 characters for a global reference; e.g. the global name+total length of keys and commas must be less than 511 characters), but that seems like a loving stupid way to do it since you're generally looping over children of a single key at any given time so you'd want to be able to have each key be a leaf that then points to a b+ tree that is the index of its children. the cache documentation claims that it never has to rebuild an index no matter how many inserts and deletes you make to the global and there is no performance hit from that, but i assume that they are relying on some semantic playing of index not referring to the internal nodes of the tree, which are, inherently, merely indices of their children. that or they're just copying and chucking a node any time they insert into it or something and they just never propagate splits up the tree or something, but again that seems like it would degrade over time. i guess this is why im not a systems engineer yet
|
# ? Nov 15, 2016 04:15 |
|
LeftistMuslimObama posted:i guess this is why im not a systems engineer yet which seems more likely, that a terrible implementation of an archaic-when-designed language like MUMPS might not always make the best choices, or that you might actually be reasonably qualified to do something I know which I'd bet on (hint she's LMO)
|
# ? Nov 15, 2016 04:47 |
|
eschaton posted:which seems more likely, that a terrible implementation of an archaic-when-designed language like MUMPS might not always make the best choices, or that you might actually be reasonably qualified to do something how are these mutually exclusive?
|
# ? Nov 15, 2016 05:26 |
|
leper khan posted:how are these mutually exclusive? eschaton is negging me in the vain hope that im not gay
|
# ? Nov 15, 2016 05:28 |
|
antlr looks so much better than jcup. gonna be messy figuring out how to deal with the weird syntax of for loops in mumps though. the $o for loop is actually 3 separate statements executing together. i guess treating it as a do-while loop with the $o call hoisted into the beginning of the loop and the quit conditions treated as conditional breaks? the scoping would be 100% the same, but i can't think of a case where it would actually change a program's behavior. hmmmmmmmm
|
# ? Nov 15, 2016 05:40 |
|
roll yer own parser
|
# ? Nov 15, 2016 05:43 |
|
I'm currently wrapping a bunch of Amiga UI crap in a homemade C++ library for some reason, probably
|
# ? Nov 15, 2016 05:46 |
|
Bloody posted:roll yer own parser gently caress that dude. there's like 0 reason to write a worse parser than the 15 existing open-source parsers written by smart autists with singleminded focus
|
# ? Nov 15, 2016 05:53 |
|
LeftistMuslimObama posted:gently caress that dude. there's like 0 reason to write a worse
|
# ? Nov 15, 2016 06:13 |
|
Luigi Thirty posted:I'm currently wrapping a bunch of Amiga UI crap in a homemade C++ library for some reason, probably why not use E? er, Amiga E? its author has this to say about it: wouter posted:I wrote the entire thing in assembly language, which will always persist in my mind as the craziest software engineering exercise I have ever gone through (400kb of mostly uncommented assembly in a single source file, for a complex compiler). eschaton fucked around with this message at 06:31 on Nov 15, 2016 |
# ? Nov 15, 2016 06:26 |
|
LeftistMuslimObama posted:learning typescript now. its better, but i have to say that it seems odd to me that theyd write a language specifically to take away the weirdness of js but then still hold onto stuff like the weird usage of colons in function signatures and poo poo like that. if youre going to make a good language that transpiles to js, why only take me half way in the name of preserving aesthetic similarity to js? imo typescript syntax has some advantages over c#, especially the type declarations. "foo: string" beats "string foo" any day and the latter only exists to not scare the java programmers.
|
# ? Nov 15, 2016 06:58 |
|
jony neuemonic posted:imo typescript syntax has some advantages over c#, especially the type declarations. "foo: string" beats "string foo" any day and the latter only exists to not scare the java programmers. var foo
|
# ? Nov 15, 2016 11:37 |
|
jony neuemonic posted:imo typescript syntax has some advantages over c#, especially the type declarations. "foo: string" beats "string foo" any day and the latter only exists to not scare the java programmers. but what about tabs/spaces?!?
|
# ? Nov 15, 2016 11:53 |
|
eschaton posted:why not use E? er, Amiga E? uhhhhhhhh well at least I'm not that insane I'm using SAS/C which is a descendant of Lattice C and the most popular compiler for classic Amiga. it's apparently still sold for IBM mainframes and nothing else. I'm sure if I poke around aminet I can find someone who did this 20 years ago other options include vbcc or GCC 2.x
|
# ? Nov 15, 2016 15:22 |
|
eschaton posted:why not use E? er, Amiga E? i wrote a few small things in E about 20 years ago. from what i remember its basically just pascal, but with more C-like syntax and a thin wrapper over some of the basic OS functions
|
# ? Nov 15, 2016 15:41 |
|
LeftistMuslimObama posted:learning typescript now. its better, but i have to say that it seems odd to me that theyd write a language specifically to take away the weirdness of js but then still hold onto stuff like the weird usage of colons in function signatures and poo poo like that. if youre going to make a good language that transpiles to js, why only take me half way in the name of preserving aesthetic similarity to js? the syntax was never the problem with js.
|
# ? Nov 15, 2016 15:43 |
|
except automatic semicolons
|
# ? Nov 15, 2016 15:43 |
|
typescript is legit. i love that you can make non-nullable default and vs code will give me red squigglies if i do something bad. the only bad thing about TS is it's compile-time only and provides no runtime features at all. like you can't have an interface with const strings like you can in java, and they refuse to implement an elvis operator because JS doesn't have one. so i and a coworker are basically in limbo due to mgmt changes - we will take over an outsourced app in a couple weeks and currently have like 0-1 hr of work a day until then, so we're just rewriting the outsourced app piecemeal so we can just throw it the gently caress out (theres still worse stuff ive found in addition to what i posted about already) we're rewriting in react, redux, and typescript and its amazing. theres a loss of type information between a couple of the pieces but besides that it's a really great environment
|
# ? Nov 15, 2016 16:07 |
|
between typescript+vscode, tslint, react storybook, and redux time travel debugging and state transition logs, theres so many safety nets and so muchconstant feedback that i almost forget im writing jabascript
|
# ? Nov 15, 2016 16:12 |
|
fleshweasel posted:typescript is meant to be a superset of JavaScript to enable gradual migration and conversion of existing projects. it's a good language that has a lot of use in the Real World what happens when microsoft looses interest in the project?
|
# ? Nov 15, 2016 16:15 |
|
is it going to die a slow painful death like haxe?
|
# ? Nov 15, 2016 16:16 |
|
then you delete the type annotations from your ts and rename to js
|
# ? Nov 15, 2016 16:18 |
|
SQL code:
|
# ? Nov 15, 2016 16:19 |
|
go play outside Skyler posted:is it going to die a slow painful death like haxe? haxe is dead? good. their just deserts for choosing to write the compiler in ocaml to make it as little accessible as possible
|
# ? Nov 15, 2016 16:35 |
|
AllScripts dudes in the office 2day training on their API. its soap but theres a single method where u specify the name of an action and then a list of optional parameters based on the action. amazing.
|
# ? Nov 15, 2016 17:06 |
|
when you want to make a rest api but your boss insists you use soap e: wait, doesn't that setup come out to be literally the worst of both worlds? HoboMan fucked around with this message at 17:32 on Nov 15, 2016 |
# ? Nov 15, 2016 17:27 |
|
http://aminet.net/package/dev/gcc/ADE I have no idea how old GCC 2.95 is but there it is
|
# ? Nov 15, 2016 17:40 |
|
HoboMan posted:when you want to make a rest api but your boss insists you use soap Indeed. Unless you love soap envelopes, I guess.
|
# ? Nov 15, 2016 17:41 |
|
HoboMan posted:
wtf now my brain hurts
|
# ? Nov 15, 2016 17:51 |
|
I think this is what I was looking forquote:This is the NEW version 1.10 by Aug, 30, 1994 of A++ Library. Looks like an object-oriented version of the Amiga libraries, woohoo
|
# ? Nov 15, 2016 17:51 |
|
Luigi Thirty posted:I think this is what I was looking for what about boopsi or whatever it is?
|
# ? Nov 15, 2016 18:29 |
|
that's for intuition, this is wrappers for the other libraries
|
# ? Nov 15, 2016 18:33 |
|
wrappers is shims with libs inside
|
# ? Nov 15, 2016 18:34 |
|
|
# ? May 9, 2024 21:45 |
|
JawnV6 posted:wrappers is shims with libs inside
|
# ? Nov 15, 2016 18:36 |